Modeling Ovarian Aging and Germ Cell Depletion

We fitted 20 asymmetric peak models to the data set, using TableCurve-2D (Systat Software Inc., San Jose, California, USA), and ranked by correlation coefficient. Each model defines a generic type of curve and has parameters, which, when instantiated gives a specific curve of that type. For each type of curve, we calculated values for the parameters that maximise the for that model. The 20 models supplied by TableCurve are those that are commonly reported in the scientific literature as models of datasets that rise and fall such as pharmacodymanics, cell populations and electromagnetic signals. The Levenberg-Marquardt non-linear curve fitting algorithm was used, with convergence to 15 significant figures in after a maximum of 10,000 iterations. We performed the same analysis on the datapoints associated with an age of 25 years or under. File S1 contains the dataset, the model ranking, the output used to prepare Figures 1 and 2, and the statistics associated with the highest-ranked model.
To avoid selection bias, we randomly removed 50 datapoints 61 times and re-fitted the models, calculating the mean and standard deviations of the coefficients obtained for each model. We then compared the mean obtained for the two highest ranked models for a statistically significant difference. File S2 contains statistics regarding the correlation coefficient for the models obtained in this way, together with output for the test for a statistically significant difference of the two highest means. To further avoid selection bias from our initial choice of models, and to allow the possibility of more than one peak (i.e. to allow models involving regeneration of ovarian reserve), we fitted all 266 models supplied by TableCurve, again ranking by . The highest ranked model was used as the basis for further calculations. Under the modelling assumption that, in general, a high (versus low) established population results in a late (versus early) menopause, we calculated the percentage of NGF pool at given ages, and the absolute monthly loss of germ cells from birth until age 55.